Google Trends With Python using PyTrends

# connect to google 

from pytrends.request import TrendReq

pytrends = TrendReq(hl='en-US', tz=360) #hl is a language selection, tz is used for timezone

build payload

kw_list = ["machine learning"] # list of keywords to get data 
pytrends.build_payload(kw_list, cat=0, timeframe='today 12-m') 

Build Payload

Build Payload is a function which is used to find the trends of all the words we enter

# build payload
# we will pass any word in the kw_list to see the search of the word with time
# For Example we are passing the word 'machine learning'

kw_list = ["machine learning"] # list of keywords to get data 

pytrends.build_payload(kw_list, cat=0, timeframe='today 12-m') 

#1 Interest over Time
data = pytrends.interest_over_time() # this function will get all the historical data of the entered word
data = data.reset_index() 


import plotly.express as px

fig = px.line(data, x="date", y=['machine learning'], title='Keyword Web Search Interest Over Time')
fig.show() 

We can pass more than one keyword so see the comparison for example we are passing the following keyword to see their comparison

• machine learning
• data science
• deep learning

kw_list = ["machine learning","data science","deep learning"] 
pytrends.build_payload(kw_list, cat=0, timeframe='today 12-m')

data = pytrends.interest_over_time() # this function will get all the historical data of the entered word
data = data.reset_index() 


import plotly.express as px

fig = px.line(data, x="date", y=["machine learning","data science","deep learning"] , title='Keyword Web Search Interest Over Time')
fig.show() 

as you can see that the deep learning is out of competition, the real competition is between machine learning and data science

Similarly lets check some major companies

kw_list = ['apple','samsung','hawaii','vivo'] 
pytrends.build_payload(kw_list, cat=0, timeframe='today 12-m')
data = pytrends.interest_over_time() # this function will get all the historical data of the entered word
data = data.reset_index() 
fig = px.line(data, x="date", y= kw_list , title='Keyword Web Search Interest Over Time')
fig.show() 

As you can see the Apple and Samsung are the best competitors, Hawaii is somehow trying to compete but Vivo is out of it

Now let's give some comparison of Bollywood Actors

kw_list = ['Shah Rukh Khan','Salman Khan','Akshay Kumar','Aamir Khan','Hrithik Roshan'] 
pytrends.build_payload(kw_list, cat=0, timeframe='today 12-m')
data = pytrends.interest_over_time() # this function will get all the historical data of the entered word
data = data.reset_index() 
fig = px.line(data, x="date", y= kw_list , title='Keyword Web Search Interest Over Time')
fig.show() 

wooo, I was thinking that Shah Rukh Khan will be on top, but in jan 2022 Salman Khan was on trending

Historical Hourly Interest

If you are interested in the hourly interest of the keyword, you can use the get_historical_intereset() method to fetch hourly data according to the time you have specified.

pytrends.get_historical_interest(kw_list, year_start=2021, month_start=9, day_start=1, 
                                 hour_start=0, year_end=2021, month_end=9, 
                                 day_end=30, hour_end=0, cat=0, sleep=0)

	Shah Rukh Khan	Salman Khan	Akshay Kumar	Aamir Khan	Hrithik Roshan	isPartial
date
2021-09-01 00:00:00	0	4	11	0	4	False
2021-09-01 01:00:00	3	13	5	3	3	False
2021-09-01 02:00:00	7	13	12	5	3	False
2021-09-01 03:00:00	3	24	17	3	13	False
2021-09-01 04:00:00	6	17	15	6	16	False
...	...	...	...	...	...	...
2021-09-29 20:00:00	0	14	9	4	9	False
2021-09-29 21:00:00	5	9	6	5	3	False
2021-09-29 22:00:00	2	10	13	4	12	False
2021-09-29 23:00:00	4	16	7	4	8	False
2021-09-30 00:00:00	4	18	6	4	4	False

701 rows × 6 columns

What does the above data frame tells us? it tells us the hourly search of these actors for example on 2021-09-01 02:00:00, shah rukh khan was searched 7 times, similarly salman khan was searched 13 times in an hour

Interest by Region

Now I want see the search results country wise

by_region = pytrends.interest_by_region(resolution='COUNTRY', inc_low_vol=True, inc_geo_code=False)

by_region.head(10) 

	Shah Rukh Khan	Salman Khan	Akshay Kumar	Aamir Khan	Hrithik Roshan
geoName
Afghanistan	10	69	10	11	0
Albania	0	0	0	0	0
Algeria	0	0	0	0	100
American Samoa	0	0	0	0	0
Andorra	0	0	0	0	0
Angola	0	100	0	0	0
Anguilla	0	0	0	0	0
Antarctica	0	0	0	0	0
Antigua & Barbuda	0	0	0	0	0
Argentina	0	100	0	0	0

The above data shows the searches by country wise,

• the above code will fetch the results from every country
• lets apply some restrictions on the dataframe
• to show results greater than 40

# by region greater than 10 searches
by_region[by_region["Shah Rukh Khan"] > 40]

	Shah Rukh Khan	Salman Khan	Akshay Kumar	Aamir Khan	Hrithik Roshan
geoName
Belgium	66	18	0	0	16
Canada	41	25	18	8	8
China	100	0	0	0	0
Costa Rica	100	0	0	0	0
Croatia	100	0	0	0	0
Egypt	46	8	15	0	31
Finland	50	0	50	0	0
Georgia	50	0	0	0	50
Guyana	67	33	0	0	0
Indonesia	54	19	9	7	11
Kenya	71	16	5	0	8
Malaysia	62	21	6	4	7
Maldives	60	20	0	20	0
Philippines	53	9	4	34	0
Réunion	100	0	0	0	0
Senegal	100	0	0	0	0
Singapore	57	19	11	5	8
Somalia	50	17	33	0	0
Spain	60	21	19	0	0
Sri Lanka	48	18	7	11	16
Taiwan	50	0	0	50	0
Thailand	65	18	11	6	0
Trinidad & Tobago	62	7	7	8	16
Tunisia	67	0	0	33	0
Ukraine	100	0	0	0	0
Venezuela	50	0	50	0	0

Related Queries

Pytrends can also help you find keywords that are closely tied to a primary keyword of your choice and then return a list of related keywords shown on Google Trends. Let us find a list of related queries for “machine learning” and return the top queries.

data  = pytrends.related_queries()

data['Salman Khan']['top'] 

	query	value
0	salman khan movie	100
1	shahrukh khan	74
2	salman khan age	59
3	salman khan movies	44
4	bigg boss	43
5	salman khan song	37
6	salman khan net worth	31
7	aryan khan	30
8	akshay kumar	30
9	salman khan height	27
10	katrina kaif	25
11	bigg boss 15	24
12	salman khan wife	24
13	salman khan photo	24
14	aamir khan	23
15	salman khan news	23
16	salman yusuff khan	22
17	amitabh bachchan	20
18	srk	17
19	salman khan height in feet	16
20	wanted salman khan	16
21	salman khan son	16
22	hrithik roshan	15
23	salman khan instagram	14
24	salman khan new movie	14

The above are the related queries and keywords searched about salman khan

• let check for Shah Rukh Khan

data['Shah Rukh Khan']['top'] 

	query	value
0	shah rukh khan son	100
1	aryan khan	35
2	shahrukh khan	14
3	shah rukh khan movies	13
4	shah rukh khan age	13
5	india	12
6	shah rukh khan son drugs	10
7	shah rukh khan net worth	9
8	gauri khan	9
9	shah rukh khan son age	8
10	salman khan	8
11	shah rukh khan wife	7
12	shah rukh khan daughter	6
13	shah rukh khan son arrested	5
14	shah rukh khan family	5
15	anak shah rukh khan	5
16	srk	5
17	shah rukh khan son news	5
18	suhana khan	5
19	आर्यन खान	4
20	shahrukh khan son	4
21	shah rukh khan children	4
22	ncb	4
23	aryan khan age	4
24	kareena kapoor	4

data['Aamir Khan']['top'] 

	query	value
0	salman khan	100
1	shahrukh khan	87
2	aamir khan movies	86
3	amir khan	70
4	aamir khan wife	60
5	aamir khan height	51
6	akshay kumar	49
7	aamir khan age	49
8	aamir khan filmleri	47
9	samantha	39
10	aamir khan actor	37
11	hrithik roshan	37
12	aamir khan height in feet	36
13	dangal	32
14	aamir khan net worth	29
15	naga chaitanya	29
16	aamir khan son	27
17	aamir khan daughter	27
18	aamir khan divorce	25
19	saif ali khan	25
20	amitabh bachchan	23
21	aamir khan wife name	21
22	athar aamir khan	19
23	salman khan net worth	18
24	shah rukh khan	18

Lets give some comparison of Google itself, Youtube, Facebook, Instagram and twitter

kw_list = ['google','youtube','facebook','instagram','twitter'] 
pytrends.build_payload(kw_list, cat=0, timeframe='today 12-m')
data = pytrends.interest_over_time() # this function will get all the historical data of the entered word
data = data.reset_index() 
fig = px.line(data, x="date", y= kw_list , title='Keyword Web Search Interest Over Time')
fig.show() 

Insta and twitter are below the line and Google is on top